Most drive-health tools report dozens of attributes with a green tick. The green tick is close to meaningless: a drive can pass its overall SMART status while carrying the specific values that predict imminent failure.
The overall SMART verdict only fails once a threshold is crossed, often far too late. SMART can alert you to problems long before most manufacturers’ thresholds are reached, but only when you’re looking at all the right metrics.
Reallocated sector count (05): those sectors have been removed from use due to reliability issues. Any non-zero value is worth your attention; if this number increments it is time to back up now. Current pending sector count (C5): this is how many unreadable sectors the drive has identified but not yet removed from use. This is your single most important early warning indicator and it often shows a non-zero value when other status readings are still reporting healthy sectors. Uncorrectable sector count (C6): how many times you got an error reading data. Reported uncorrectable errors (BB), and the spin retry count (0A), make up the rest of them; a rising spin retry count indicates possible problems with either the motor or bearings.
Every other variable (such as temperature, number of hours your device has been powered on, and number of load cycles) is considered in terms of being part of the context, not predictive information.
When viewing your disk attributes using Windows-based systems, CrystalDiskInfo is free and will show all of your disk attributes in their raw value form. The manufacturer-created applications (i.e., SeaTools and WD Dashboard) provide an additional layer of interpretation based upon the vendor’s specific criteria. On Apple-based systems, System Information provides the basic status of each disk but does not include detailed attributes. Smartmontools can be used on both Apple- and Linux-based systems to view detailed information about each disk. To do this, simply type “smartctl -a /dev/sdX” from within the terminal where X represents the target disk.
When reviewing your disk attributes, focus on viewing them in their raw format as opposed to their normalised formats. Disk manufacturers normalise many of their critical attributes by relating them to their own threshold criteria. This can allow a disk to pass its general health test while the underlying raw counts continue to rise and accumulate potential bad sectors.
If you hear a clicking sound coming from your hard disk approximately every couple of seconds while it is running, then it is likely that the disk head(s) have lost their ability to read/locate servo data. You should immediately turn off the system. Grinding or scraping sounds emanating from the disk when it is spinning indicate that one or more heads has made contact with the platters’ surfaces, resulting in permanent damage for each passing second. Buzzing noises that occur repeatedly without evidence of a spin up may indicate either stiction or a seized motor. High pitched squealing sounds, where none existed previously, typically indicate bearing failure.
There is no software solution available for any of these issues. As such they are going to get progressively worse as you continue to use them. Any drive that makes an unusual noise should immediately be removed and checked out, not tested.
If you see either pending or reallocated sectors increasing on a drive, remove (copy) all important data first before proceeding with anything else, not after you have run full surface scans on the drive that has just told you it is having problems. Do not try to fix the drive; replace it instead. A drive which has started retiring sectors will continue to do so.
In circumstances where the drive has already failed, recovery can cost from £300 +VAT per single drive after a complimentary 48-hour diagnostic process; however, in most cases you only pay if we recover your data.
To ensure accuracy do not rely solely upon the overall health of the drive (SMART status), but read each SMART attribute individually: CrystalDiskInfo (Windows) / smartctl (macOS & Linux). The only SMART attributes that truly matter are reallocated sectors, current pending sectors, uncorrectable sectors, reported uncorrectable errors and spin retry count.
Not always. SMART reports an overall “failure” only after a manufacturer-set failure-threshold (which is usually quite high) is exceeded, and that is often very late. The drive may have reported itself as healthy with an increasing number of pending sectors; similarly, slow performance typically results from a drive’s repeated attempts to read bad sectors.
Bad sectors that are currently known to the drive but have been neither removed nor retired. This number will often rise before the overall SMART status indicates a problem. This is your earliest possible indicator of trouble, and if this number rises at all, back up immediately.
Not before backing up. On a drive that is making noise, don’t even think about it. If you perform a surface test on a failing drive you risk causing additional problems with the mechanical components. Once the drive starts showing signs of mechanical failure, it should never again be asked to do a full surface test.
Single drives (laptops/desktops/hard drives etc.) can be restored from £300 + VAT, with a firm price given in a written quotation following a complimentary 48-hour diagnostic. In most cases you only pay if we recover your data; however, should we agree to undertake ‘donor part’ work, this would incur a 50% deposit payment prior to commencement of such work. Only upon successful retrieval of your data will the balance be payable.